Depth of anesthesia prediction via EEG signals using convolutional neural network and ensemble empirical mode decomposition

According to a recently conducted survey on surgical complication mortality rate, 47% of such cases are due anesthetics overdose. This indicates that there is an urgent need moderate the level anesthesia. Recently deep learning (DL) methods have played major role in estimating depth Anesthesia (DOA) patients and has essential control anesthesia In this paper, Electroencephalography (EEG) signals been used for prediction DOA. EEG very complex which may require months training advanced signal processing techniques. It point debate whether DL improvement over already existing traditional approaches. One algorithms Convolutional neural network (CNN) popular algorithm object recognition widely growing its applications hierarchy human visual system. various decomposition extracting features signal. After acquiring necessary values image format, several CNN models deployed classification DOA depending upon their Bispectral Index (BIS) quality index (SQI). The were converted into frequency domain using Empirical Mode Decomposition (EMD), Ensemble (EEMD). However, because inter mode mixing observed EMD method; EEMD utilized study. developed predict based spectrum images without use handcrafted provides intuitive mapping with high efficiency reliability. best trained model gives accuracy 83.2%. Hence, further scope research can be carried out methods.